Hamim Touria

PhD in Computer Science / Data Science with a sustained publication record in machine learning applied to educational data mining and decision-support systems, combined with 6+ years leading applied AI/ML in industry (telecom and large-scale retail). Experienced in supervising and evaluating final-year engineering projects, designing pedagogical material, and bridging academic methodology with production-grade machine learning (MLOps, cloud/Databricks, LLM applications, geospatial and predictive modeling). Seeking a faculty or applied-research position in Oman where this combination of research grounding and industry delivery can support both teaching and real-world-oriented research.

Tech Lead Data Science & AI at LabelVie Groupe (2025-08 – Present)

• Led and mentored a team of 5 data scientists (planning, code reviews, upskilling, performance follow-up).
• Designed and enforced the Data Science Governance Framework: Git branching strategy, project templates, MLflow standards, production guidelines.
• Drove the full migration of Data Science environments to Azure Databricks (Unity Catalog, dedicated ML clusters, DEV/PROD catalogs, jobs orchestration).
• Implemented MLflow pipelines, experiment tracking, model registry, monitoring and cost-optimization strategies.
• Partnered closely with Data Engineering, DevOps and Data Platform teams on architecture, data quality, security and workload orchestration.
• Customer Churn Prediction: scoring model, decile segmentation, business uplift analysis.
• Retail Catchment: geospatial scoring engine to guide store-expansion decisions across 6 retail formats, supporting LabelVie's 2028 growth plan; deployed as an interactive Kepler.gl/Streamlit app on Databricks Apps.
• Product Boost: product-recommendation models delivering high-ROI uplift.
• Fraud Detection: hybrid rules + deep-learning autoencoder for cashier/loyalty fraud.
• RFM Segmentation & Potential Scoring: customer clustering, value scoring, business dashboards.
• E-Reputation Analysis: large-scale scraping + LLM sentiment analysis on 600k+ Google Maps reviews.
• LabelBot: internal chatbot on Databricks (Teams/WhatsApp) for natural-language sales insights.
• LIVIA – AI Validation Assistant: intelligent email validator ensuring automatic data-quality checks for business teams.

Data Scientist at Orange (2020-03 – 2025-06)

• Built a random-forest churn model (SAS Guide, Python, PySpark, GCP) achieving a 96% improvement in B2C churn-prediction accuracy, capturing 40% of churn by targeting only 10% of customers.
• Engineered a predictive model targeting mobile B2C postpaid customers with minimal top-up amounts while favoring additional top-ups (Python, SAS), averaging 85% prediction accuracy.
• Built random-forest and LightGBM models for intra-postpaid tariff migration (up-tariff and down-tariff transitions), reaching ~99% accuracy.
• Implemented a prediction and recommendation system to drive prepaid-to-postpaid migration, achieving 97% accuracy.
• Implemented a Khiops-based predictive model for churn and additional-recharge propensity in B2C mobile, reaching ~97% accuracy on both models.
• Led an end-to-end reporting project (data gathering, process automation, SAS Visual Analytics) for a global mobile-activity report, cutting manual effort by 80%.
• Automated and optimized data-analysis processes using Zeppelin, PySpark, SQL, SAS and Business Objects.

PhD in Computer Science / Data Science – Faculty of Sciences Ben M'Sick, Hassan II University of Casablanca (2022-03)

Engineer's Degree in Information & Communication Systems Engineering – National School of Applied Sciences (ENSAJ) (2017-09)